Smoothed Kalman Filter Pose Estimation for Mobile Machines
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing positioning systems for autonomously controlled machines experience error accumulation and discontinuities in trajectory due to the lower update rate of Kalman filters, leading to sudden changes in velocity or orientation, which can cause discomfort and safety hazards.
Innovation Solution
A positioning system that updates pose estimates using a fraction of the correction determined by a Kalman filter, spreading the correction over each propagated pose estimate to achieve a smoothed update, thereby eliminating discontinuities and ensuring smooth trajectory propagation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the Kalman filter runs at a lower rate (e.g., 1 Hz) to keep updates sufficiently separated in time and eliminate time correlated errors, then measurement precision is improved, but the trajectory smoothness deteriorates due to noticeable error accumulation and discontinuities
Solution Approach 1:
The correction determined by the Kalman filter is segmented into multiple fractional corrections, each applied to intermediate pose estimates between GPS updates. This divides the single large correction into smaller incremental adjustments, maintaining measurement precision while ensuring trajectory smoothness.
Solution Approach 2:
The system performs preliminary pose propagation at a higher rate using IMU data to generate intermediate pose estimates before applying the fractional corrections. This preliminary action ensures continuous trajectory generation while the fractional corrections are applied to maintain smooth transitions.
2Stability of the object's composition
If the Kalman filter update rate is increased to reduce error accumulation, then trajectory smoothness is improved, but time correlated errors increase affecting measurement precision
Solution Approach 1:
The system applies only a fraction of the total correction to each intermediate pose estimate rather than applying the full correction at once. This partial action approach distributes the correction effect across multiple time points, ensuring smooth trajectory propagation while maintaining the accuracy benefits of the lower-rate Kalman filter.
3Productivity
If the IMU data update rate is high (higher than 100 Hz) to provide continuous pose estimates, then productivity is improved, but sudden changes in velocity or orientation occur causing safety hazards
Solution Approach 1:
The system uses feedback from the Kalman filter correction to adjust the intermediate pose estimates generated at high IMU update rates. Each propagated pose estimate is updated with a fractional correction based on the GPS measurement, creating a feedback mechanism that ensures operational safety while maintaining high productivity.
Data Source
AI summary
A method and a system of determining a position of a mobile machine are disclosed. According to certain embodiments, the system may include a first sensor configured to generate a first signal indicative of a parameter of the mobile machine. The system may also include a second sensor configured to generate a second signal indicative of a pose of the mobile machine. The system may further include a controller in communication with the first and second sensors. The controller may be configured to generate one or more estimated poses of the mobile machine based on the first signal. The controller may further be configured to update each estimated pose with a fraction of a correction. The correction may be determined, based on the second signal, in a measurement update stage of a Kalman filter.